FormalPara Key Summary Points

Pythium insidiosum is an oomycete causing vision-threatening keratitis in humans.

Presumptive assumption by the clinician and the microbiologist that they are dealing with a fungus can delay early diagnosis and initiation of appropriate treatment.

Recent literature supports growing evidence for the use of antibacterials (linezolid and azithromycin) as the first-line drugs.

Due to high virulence, rapid proliferation, and early recurrence, early therapeutic keratoplasty with 1-mm margin clearance is recommended.

Introduction

The genus Pythium embraces 200 species. They are one of the most destructive plant pathogens, infecting crops and destroying the seeds, storage organs, roots, and other plant tissues [1]. The genus was considered as true fungi by Pringsheim in 1858 due to characteristics including eukaryotes with filamentous growth, absence of septa, nutrition by absorption, and reproduction via spores [2]. However, as the understanding of evolution has developed, the Pythium species are now being classified as protozoan, as they have cellulose, beta-glucans, and hydroxyproline in their cell walls [2]. The only species of Pythium known to cause keratitis in humans is Pythium insidiosum (PI) [3], which was described as an emerging phytopathogen causing infections in animals and humans in 1884 [4]. Virgile et al. made the first report of a corneal ulcer in a 31-year-old woman caused by PI in 1993. The patient was subsequently diagnosed with sickle cell trait [3]. Since then, 168 ocular pythiosis cases have been reported in the literature until 2021 [5]. PI keratitis has been sparsely reported due to varied reasons such as lack of knowledge about the microscopic and colony morphology of the organisms among the microbiologists, falsely labeling it as an unidentified fungus, absence of growth in the culture, lack of high clinical suspicion among the clinicians, delayed diagnosis, and its clinical resemblance to fungus [6]. The common systemic and ocular risk factors identified for Pythium infection are thalassemias, sickle cell trait, swimming with contact lenses, exposure to farm animals, and swampy areas [7]. Recently, a few studies from South India reported a higher incidence of PI keratitis in homemakers and IT professionals [8, 9].

Early diagnosis and prompt differentiation from fungal keratitis (FK) are imperative for immediate initiation of appropriate treatment for Pythium to yield a good prognosis [7]. The direct microscopic examination of smears with Gram staining and KOH mount, on the cursory look, may resemble fungal filaments. The sensitivity ranges from 79.3 to 96.5% [8]. Still, the microbiologist's familiarity with the typical appearance of Pythium filaments, which appear as broad, ribbon-like filaments with right angle branches, would help in the early identification of the infection [9]. The confirmatory test is the cultural identification with zoospore formation but it takes 5–7 days, which may significantly delay the initiation of treatment for Pythium [8, 10, 11]. The time to identification can be minimized if the clinician and microbiologist are experienced and suspect Pythium. Agarwal et al. had used confocal microscopy to rapidly identify and define the typical confocal features of Pythium in the stroma; they appear as multiple, linear, hyper-reflective, well-delineated structures with a width of 4 μm and length of 350 μm seen in all the layers of the cornea with occasional branching and intersecting angles at 78.6° [12]. Though it has a 95% sensitivity, it cannot differentiate PI from other FK [12, 13]. Serodiagnosis to detect antibodies against P. insidiosum, immunofluorescence, or immunoperoxidase staining assays can assist in diagnosis if a high index of clinical intuition is present [9]. Polymerase chain reaction (PCR) is considered the gold standard in the diagnosis of Pythium. It is crucial to understand the recent diagnostic modalities for prompt identification of the PI, with high specificity and sensitivity for the expeditious management [14]. Previously, therapeutic keratoplasty remained the terminal option to reduce the microbial load to salvage the eye and invariably 90% of the keratitis ended up with evisceration [15]. Recently, several authors have reported in vitro susceptibility of various strains to antibacterial such as tigecycline, macrolides, tetracyclines, and linezolid both as monotherapy and in combination with antifungal agents [10, 11, 14, 16,17,18].

However, with better understanding of its morphology, clinical features, and response to antibacterial agents along with the development of adjunctive treatment methods, the prognosis and outcomes have significantly improved [16,17,18]. This review article discusses the entire evolution of our understanding of PI keratitis concerning its morphology, clinical features, laboratory diagnosis, and management in detail along with a futuristic perspective on the way forward. This work is based on previously conducted studies, and does not contain any new studies with human participants or animals performed by any of the authors.

Method of Literature Search

A detailed systematic literature search was performed of the PubMed, Google Scholar, ePub publications, and Cochrane Library database for all the recent case reports, series, original articles, review articles, and clinical trials on PI keratitis. The literature search was performed using the keywords Pythium insidiosum, Pythium insidiosum keratitis, Pythium keratitis, Pythium AND (Zoospore) AND (antifungals) AND (antibacterials) AND (review) AND (diagnosis) AND (treatment). Articles in languages other than English were also reviewed. All the relevant articles were compiled and reviewed for relevant literature.

Epidemiology

Pythium insidiosum keratitis is reported commonly in tropical, subtropical, and temperate climates [10]. The first case of ocular pythiosis came into existence in 1988 from Thailand. Since then, there have been numerous reports from different countries, viz. Australia, Israel, China, Japan, Malaysia, India, etc. [7]. As per detailed literature review, the reported incidence and prevalence of PI keratitis is variable, which is probably due to the rarity of microorganisms and regional and geographical variations. Das et al. reported in their analysis of keratitis in COVID-19 period reported it to be a rare cause of keratitis compared to FK [19]. Hasika et al. reported a prevalence of 5.9% (71/1204) of PI keratitis from South India [15]. Vishwakarma et al. in their analysis reported a higher male preponderance and greater prevalence among agricultural workers of lower socio-economic strata [20]. Sharma et al. reported a prevalence of 5.5% (9/162) in phase 1 and 3.9% (4/102) in phase 2 of PI keratitis as an etiological agent of FK [11]. A higher incidence of PI has been reported in patients with thalassemia, paroxysmal nocturnal hemoglobinuria (PNH), chronic arterial insufficiency and patients with thrombosis, aneurysms, and vasculitis [7]. Hence, a high index of clinical suspicion is necessary among clinicians to diagnose the vision-threatening keratitis [21].

Clinical Features

Clinical features of PI keratitis can be highly variable, which makes the diagnosis challenging. Similar to FK, the corneal surface overlying the infection is described as "dry" without the significant suppuration or purulence that is associated with bacterial keratitis [16]. PI keratitis can present with several patterns of infiltration. Across several case series, the most common presentation described is a dense, full-thickness stromal infiltrate in the central cornea with associated radiations of tentacle-like, wispy, reticular infiltrates in the subepithelial layer or the anterior corneal stroma [12, 16, 22,23,24,25]. Despite being distinctive, these radiating reticular infiltrates are neither pathognomonic nor present in all cases [11]. PI keratitis can also present with a large central infiltrate associated with multiple smaller discrete satellite infiltrates with intervening gaps of the clear cornea [11, 16, 24]; multifocal infiltrates scattered across the cornea [11, 16]; punctate miliary subepithelial lesions [10, 17]; or a large, dense infiltrate occupying most of the cornea [16, 24, 25]. Infiltrates often have feathery or non-discrete edges, similar to FK [17]. Other non-specific findings, including ring infiltrates [26], keratoneuritis [23], keratic precipitates [26], endothelial plaque [10, 27], and peripheral corneal thinning [10], have also been reported. Hypopyon has been commonly reported, particularly in severe cases [11, 12, 22, 25, 27]. In severe cases, the infection can extend to the limbus, sclera, and anterior chamber [10, 25] or progress to corneal melt or perforation (Fig. 1) [10, 15].

Fig. 1
figure 1

Digital image of a patient with rapidly proliferative Pythium insidiosum keratitis. a At presentation (day 1)—5 × 6 mm central full-thickness infiltrate with trace hypopyon. b, c (day 7) Worsening of full-thickness infiltrate with rapid spread towards limbus and increase size and density of hypopyon despite topical medications. d Recurrence-graft infection noted 7 days following therapeutic penetrating keratoplasty, e 1 month following a regraft-diffuse congestion, stromal edema, and 360-degree superficial vascularization

Full size image

PI keratitis can most easily be misdiagnosed as FK [11, 28] due to similar clinical features, including a “dry” corneal surface and feathery infiltrates with indistinct edges, multifocal infiltrates, and lack of response to the usual antibacterial therapy. Both Pythium and FK occur more commonly among young, working-age adults with exposure to vegetable matter [10] or natural bodies of water [23, 25]. FK in developing countries classically occurs in farmers and day laborers, while multiple case series report PI keratitis in white-collar workers with no discernable exposure to plant or vegetable matter [10, 23]. Clinicians, therefore, should maintain a high index of suspicion for Pythium in cases of culture-negative or smear-negative presumed FK cases that fail to respond to empiric antifungal treatment [23]; atypical-appearing keratitis that occurs following exposure to natural bodies of water, particularly in tropical settings such as Thailand or India; and the appearance of distinctive reticular or tentacle-like superficial infiltrates suggestive of PI infection.

Acanthamoeba keratitis can also share features with PI keratitis, including ring infiltrates [26], multifocal infiltrates [16, 24], and keratoneuritis [23]. Although Acanthamoeba keratitis is most commonly seen in contact lens wearers [29], and PI keratitis is most strongly associated with exposure to natural water, there can be overlap in risk factors. Acanthamoeba is a free-living protist found particularly in aquatic environments and can cause keratitis in non-contact lens wearers, particularly in India, where the leading risk factor is exposure to vegetable matter [30, 31]. PI keratitis has been reported in contact lens wearers [23], typically after exposure to natural water [27]. Co-infection with Pythium and Acanthamoeba has also been reported [29]. To improve the diagnosis of PI keratitis, treating clinicians should maintain an open mind about the etiology of any presumed microbial keratitis that is failing to respond to empiric antimicrobial therapy. Repeat smears, molecular testing such as PCR, and/or biopsy with appropriate stains and culture should be strongly considered in such cases. A summary of the clinical features of PI keratitis and its differential diagnoses is given in Table 1 [7,8,9,10,11,12, 15, 17, 29, 32, 33].

Table 1 Classical clinical features of Pythium insidiosum keratitis and resemblance to other keratitis
Full size table

Microbiological Laboratory Diagnosis

The General Approach to Lab Diagnosis

PI keratitis, as we understand it, is relatively rare. Still, clinicians and microbiologists should always have a high suspicion index whenever dealing with atypical microbial keratitis, as missing the diagnosis usually relates to poorer outcomes [32]. Clinicians should ideally not rule out Pythium based on one form of testing alone until attaining a satisfactory clinical endpoint, as it may require multiple and/or different types of specimens ranging from a corneal scrape, corneal biopsy, corneal button to eviscerated tissue to establish the diagnosis [15, 34]. In general, any specimen requiring testing for Pythium growth should be stored between 28 and 37 °C [35]. Culture positivity with zoospore induction gives a definitive diagnosis but still PCR (polymerase chain reaction) is the gold standard due to high sensitivity and specificity. It is also vital to understand all the existing and evolving modes of lab diagnosis [36].

Direct Staining/Examination

Corneal scrapings collected under aseptic precautions can be directly stained and studied under a microscope. 'Broad sparsely septate ribbon-like hyaline filaments' is the typical description of Pythium [7]. They can also exhibit collapsed walls and vesicular expansion [37]. In contrast, fungal hyphae are broad sparsely septate with branching at various angles. While it is generally considered difficult to differentiate Pythium from fungal filaments, newer stains are helping improve the diagnostic ability of this technique. Various stains have been reported in different studies, including Gram, 10% potassium hydroxide (KOH), Calcofluor-White (CFW) [37], Acridine Orange [38], Lactophenol cotton blue (LPCB), Trypan blue, and potassium iodide-sulfuric acid (IKI-H2SO4) [9].

A combination of 10% KOH-CFW staining has been shown to have a sensitivity between 79.3 and 96.5% and over 93% specificity in detecting Pythium [37]. This requires a fluorescence microscope and microbiologist to study images. Trypan blue is an easily available stain that can be used by ophthalmologists and directly studied without any special microscope. This technique has a sensitivity of over 75% and specificity of 68% [39, 40]. IKI-H2SO4 stain is a recent addition that specifically stains Pythium, but not fungal filaments, and different studies show a specificity close to 100% [9, 41]. The sensitivity and specificity of various stains is with respect to culture results of PI keratitis. One or different combinations of the above, along with a trained eye and clinical suspicion, can help in the early detection of PI keratitis. In vivo confocal microscopy (IVCM) has been shown to demonstrate hyperreflective beaded string-like branching structures, but they cannot differentiate Pythium from fungus or Nocardia [36, 42].

Culture

Pythium has been described to grow as a flat colorless or light brown colony with fine radiations in the edges [7, 36]. Culture can be obtained either from scraping or inoculation from the corneal button into blood agar, potato dextrose agar (PDA), or Sabouraud dextrose agar (SDA). PI will grow on most commonly used bacterial, fungal, and amoebic agars (Fig. 2a–c). Grass leaves with water culture has been shown to induce zoospore formation, which is confirmatory for P. insidiosum (Fig. 2d) [11, 36]. The above process can be time-consuming (up to 11 days) and significantly delay the initiation of appropriate pharmacotherapy. Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF) is a novel technique that has been shown to rapidly identify Pythium species in the very early stages of culture. This could be used for very early detection of Pythium from culture plates with accuracy [43].

Fig. 2
figure 2

Digital culture images of Pythium insidiosum keratitis. a 5% sheep blood agar image depicting flat, gray-white colony at 37° (2nd day). b Dense cream-colored colony of Pythium on 5% sheep blood agar at 37° after 5 days. c White submerged colony on chocolate agar after 2 days. d Magnified view (× 10) of a small vesicle with numerous zoospore formation after 3 h of incubation using leaf incarnation method

Full size image

Histopathological Examination (HPE)

HPE of formalin-fixed paraffin-embedded corneal buttons excised during therapeutic keratoplasty (TPK) or eviscerated specimen is possible with the help of different stains such as H&E (hematoxylin-eosin stain), periodic acid–Schiff (PAS), and Gömöri methenamine silver (GMS) [7, 9]. This test is typically done at later stages of the disease but can help in confirming the diagnosis. IKI-H2SO4 stain typically stains Pythium but not fungal filaments and has 100% specificity, which is better than PAS or GMS stains [9].

Serology-Based Tests

Serology-based tests detect antigens or antibodies, and tests like hemagglutination (HA), ELISA, immunochromatography (ICT), and immunodiffusion (ID) are available for the detection of pythiosis [44,45,46,47]. Immunoperoxidase staining assay, ELISA, and ICT have high sensitivity and specificity [47, 48]. Overall, they have good utility in systemic pythiosis but not in keratitis, as there is usually no significant systemic antibody response in keratitis.

Molecular (PCR-Based) Tests

Apart from positive culture with zoospore formation, PCR tests are the only gold standard, as they typically identify species-specific sequences of the DNA [11, 36]. PCR can either be performed as a rapid test from a corneal scraping or the growth on a culture plate. rDNA-ITS (internal transcribed spacer) regions are usually targeted [49, 50]. PCR for Pythium detection has rapidly evolved. The initial nested PCR had a sensitivity of around 90% with 80% specificity [51]. Duplex PCR targeting 18S rRNA and ITS showed 100% specificity but 91% sensitivity [49]. The newer LAMP (loop-mediated isothermal amplification) PCR technique has 100% sensitivity and 98% specificity [52]. Real-time PCR targeting a protein exo-1, 3-beta-glucanase is a novel value addition with 100% sensitivity and specificity apart from a shortened turnaround time of 7.5 h [14]. While the above targets are used for clinical testing, PCR for targets like cox-II, ELI025, exo-1,3-β-glucanase gene, and other techniques like AFLP (amplified fragment length polymorphism) are being used for population studies of Pythium [53,54,55,56,57,58].

Imaging

Imprecise and delayed diagnosis based on different staining techniques, cultures, and zoospore identification has necessitated the need to look for alternatives. Serological tests using antibodies have facilitated the diagnosis of systemic pythiosis, but its role in ocular infection is limited [50]. The use of in vivo confocal microscopy has also been described to diagnose PI; however, the features are not characteristic or specific and cannot be used to distinguish from fungi. They appear as beaded, thin hyperreflective branching structures with a mean branching angle of 78.6° and varying from 90 to 400 μm in length [42]. Nevertheless, confocal microscopy does have a role in detecting an early recurrence [14]. The use of immunoperoxidase staining using P. insidiosum antibodies through various sources has been reported to have higher sensitivity and specificity than the routine histopathological stains; however, Kosirurkvongs et al. compared nested PCR with culture and immunostaining and found PCR to be the most sensitive [48, 59]. Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) is a recent tool for identification that is based on crude proteins generating a main spectral profile (MSP) which is then searched against a reference database for P. insidiosum [43, 60].

Treatment

Increased awareness, improved understanding of the pathogenesis, clinical presentation, and microbiological features have facilitated relatively faster diagnosis, however, management of PI keratitis still poses multiple challenges [61]. Presently, there is no gold standard treatment for PI keratitis, and with the lack of standardized antimicrobial susceptibility testing for oomycetes, choosing an appropriate antimicrobial also becomes difficult [7]. Based on broth microdilution, Etest, and disk diffusion methods following the CLSI guidelines, in vitro studies have shown tigecycline, minocycline, azithromycin, and clarithromycin to be effective as compared to antifungals [62, 63]. The lack of ergosterol in the Pythium cell wall, which is the leading site of action of all antifungals, is the primary reason for its lack of response. The Pythium cell wall primarily comprises cellulose, chitin, and small amounts of beta-glucan. Caspofungin, which inhibits beta-glucan synthase, was more effective than other echinocandins. However, the effect was only static, and thus its use was not very encouraging [64].

Loreto et al. found azithromycin to be superior to linezolid; however, Bagga et al., in their in vitro and animal studies, have found linezolid to be more effective [63, 65]. The diversity in the susceptibility techniques and the concentration of spores used possibly explain the diversity for MIC obtained for P. insidiosum. Inhibition of protein synthesis by these antibacterial agents has been considered the possible mechanism of action for inhibiting Pythium. However, bacteria are prokaryotes with 70S ribosomes with 30S and 50S subunits, whereas Pythium, like humans, are eukaryotes that possess 80S ribosomes with 60S and 40S subunits. Both linezolid and azithromycin act on the 50S ribosome unit, prevent protein synthesis, and thus act as antibacterials [66, 67]. This is in fact one of the critical factors that reduce its toxicity in humans, as it does not affect the 80S ribosome, thus raising doubts about their exact mechanism of action in inhibiting Pythium.

Recently, in a series of 69 eyes, 55.1% resolved with medical treatment, with 34.3% requiring cyanoacrylate glue for tectonic support, with the median duration of treatment being 3 months. The remaining 44.9% required a therapeutic graft. On comparing the characteristic features, eyes with infiltrates > 6 mm, longer duration, and posterior stromal involvement did not respond to medical therapy being constituted by a combination of topical linezolid, azithromycin, and oral azithromycin. The authors recommend initiating medical treatment in mild-to-moderate cases, and therapeutic keratoplasty (TPK) can be planned based on the treatment [68]. In another series of 18 eyes from Eastern India, all seven eyes where medical therapy was initiated ended up requiring a therapeutic graft or were eviscerated [28]. Among 30 eyes reported by Gurnani et al., seven (23.3%) responded to topical azithromycin and linezolid with oral azithromycin, whereas 19 eyes (63.3%) required surgical intervention. They also concluded that patients with mild-to-moderate ulcer severity with a size less than 4 mm, only anterior one-third stromal involvement, and fewer tentacular projections, respond to medical therapy. Among the 19 eyes which underwent a TPK, recurrence was noted in 20%, requiring a repeat graft [9]. Oral azithromycin can cause nausea, vomiting, abdominal pain, or stomach upset, but none has been reported by any studies. A higher incidence of recurrence in the range of 46–54.2% following TPK, even necessitating evisceration, has been a major concern in most published series [15, 24]. To reduce this incidence, a larger graft with at least 1 mm of clear margins instead of 0.5 mm is routinely indicated, including the radiating reticular pattern in the cornea within the keratoplasty (Fig. 3). Delay in presentation and diagnosis are the important parameters which govern the spread of PI to the limbus, sclera, and the presence of non-resolving anterior chamber exudates, although it is difficult to pinpoint the exact percentage or number of cases developing this picture. Moreover, once the organism has spread to the limbus, rapid scleral involvement is seen, and it becomes too difficult to contain the infection even after an aggressive keratoplasty. Full-thickness large ulcers involving the visual axis need an early keratoplasty after 7–10 days to obtain a good anatomical and functional outcome.

Fig. 3
figure 3

Digital image of a patient of Pythium insidiosum keratitis. a Depicting 6 × 8 mm full-thickness infiltrate with subepithelial infiltrates radiating up to the limbus. b No recurrence noted 2 weeks after therapeutic penetrating keratoplasty taking 1 mm larger host margin with intraoperative cryotherapy and alcohol

Full size image

Agarwal et al. proposed using surgical adjuncts like cryotherapy and ethanol in grafts involving the limbus to reduce recurrence further. The actual spread of the organism into the limbus and the adjacent sclera becomes difficult to distinguish clinically until active necrosis sets in, contributing to the increased risk of recurrence. The surgical adjuncts, both cryotherapy, and ethanol, aid in reducing recurrence following a therapeutic graft by addressing this clinically unidentifiable spread when the limbus is involved [12]. In their series of 46 eyes, recurrence rates of 100, 51.8, and 7.1% were noted in eyes that received medical treatment, therapeutic keratoplasty alone, and TPK with surgical adjuncts, respectively. A single freeze–thaw cryotherapy using a nitrous oxide probe is recommended at the limbus, and in cases with infiltration extends 3–4 mm beyond the graft–host junction, ethanol with multiple rows of cryotherapy is used to prevent recurrence [25].

Based on their initial success with ethanol, Agarwal et al. explored the efficacy of using topical ethanol to treat PI. They performed microbiological, clinical, and histopathological studies to analyze the effect of absolute ethanol on P. insidiosum along with infrared spectroscopy to assess the corneal penetration and concluded that it could be considered as a therapeutic option; however, the exact dose and strength of ethanol at which it will be most effective still needs to be explored [69]. Ethanol is known to decrease the viability of cells by causing cell lysis, suppressing proliferation, and inducing apoptosis. It primarily acts on the cell membranes, and the lack of ergosterol in the cell wall of Pythium makes it more susceptible to ethanol even at lower concentrations. Based on the reports available in the literature and the author's own experience managing PI keratitis, we propose a treatment algorithm taking the size of infiltrate at presentation as the determinant (Fig. 4).

Fig. 4
figure 4

Image depicting the proposed management algorithm for Pythium insidiosum keratitis

Full size image

Future Directions

Management of PI keratitis is challenging, and still relies on surgical technique to be the most favorable. Multiple drugs, drug combinations, natural extracts, natural compounds, bacterial isolates, substances like biogenic silver nanoparticles, copper acetate, and drugs like miltefosine have shown anti-P. insidiosum effects but need to be explored clinically [62, 70]. Xanthyletin, a natural compound of green plants, is a coumarin derivative and was found to have anti-Pythium effects by probably affecting the protein composition of the cell wall or cell membrane [71]. The presence of cellulose in the Pythium cell wall, similar to acanthamoeba cyst, makes it resilient in nature [72]. Although chemically homogenous, cellulose exists in crystalline and amorphous topologies, and there is no single enzyme that can hydrolyze it. The insoluble, crystalline, and heterogeneous nature makes it a resilient and challenging substrate for enzyme hydrolysis. Streptomyces species are being used as a biocontrol agent against Pythium species and are shown to produce enzymes that cause cell wall rupture. S. rubrolavendulae S4 was reported to have a very high cellulase activity and inhibit Pythium growth [73]. Understanding the cellulose structure and arrangement of fibers in the Pythium cell wall will help discover the most effective combination of cellulases that can disrupt the Pythium cell wall. Photodynamic therapy (PDT) is being explored to treat refractory infections, especially due to increasing antimicrobial resistance. The photosensitizer and the light interaction produce reactive oxygen species, which are highly toxic to cells. PDT using methylene blue, Photogem®, and Photodithazine® was evaluated, and all were found to have inhibition rates of more than 50% but regrowth was observed within 7 days. The three dyes' penetration varied, and Photodithiazine® showed maximum efficacy due to higher cell wall penetration and concentration in membranes [74].

Pythium has a very different cell wall structure as compared to other microorganisms. The presence of cellulose makes the drug penetration and action a challenge [75]. An improved understanding of the cell wall structure and organization will further open new avenues to be explored in managing this refractory organism. A summary of review of literature of all the clinical studies and case reports or PI keratitis is listed in Table 2 [3, 8, 10, 12, 15,16,17,18,19, 22,23,24,25,26,27,28,29, 32,33,34, 38, 68, 75,76,77,78,79,80,81,82,83,84,85].

Table 2 Review of literature of clinical studies and case reports of Pythium insidiosum keratitis
Full size table